Light-Induced Open Circuit Voltage Increase in Polymer Solar Cells

Dec 17, 2014 - We report a light-induced open circuit voltage (VOC) increase in polymer solar cells with ternary bulk heterojunction (BHJ) layers that...
0 downloads 0 Views 6MB Size
Download PDF
Research Article pubs.acs.org/journal/ascecg

Light-Induced Open Circuit Voltage Increase in Polymer Solar Cells with Ternary Bulk Heterojunction Nanolayers Sungho Nam,† Sooyong Lee,† Jaehoon Jeong,† Jooyeok Seo,† Hwajeong Kim,*,†,‡ Dong-Ik Song,§ and Youngkyoo Kim*,† †

Organic Nanoelectronics Laboratory, School of Applied Chemical Engineering, Kyungpook National University, University Road 80, Buk-gu, Daegu 702-701, Republic of Korea ‡ Research Institute of Advanced Energy Technology, Kyungpook National University, University Road 80, Buk-gu, Daegu 702-701, Republic of Korea § Polymer Rheology Laboratory, School of Applied Chemical Engineering, Kyungpook National University, University Road 80, Buk-gu, Daegu 702-701, Republic of Korea S Supporting Information *

ABSTRACT: We report a light-induced open circuit voltage (VOC) increase in polymer solar cells with ternary bulk heterojunction (BHJ) layers that are composed of poly(3-hexylthiophene) (P3HT), poly[(4,8-bis(2-ethylhexyloxy)-benzo[1,2b:4,5-b′]dithiophene)-2,6-diyl-alt-(N-2-ethylhexylthieno[3,4-c]pyrrole-4,6-dione)-2,6-diyl]] (PBDTTPD), and [6,6]-phenyl-C61butyric acid methyl ester (PC61BM). The ternary BHJ layers were prepared by varying the composition of donor polymers at a fixed ratio (1:1 by weight) of donor (P3HT + PBDTTPD) to acceptor (PC61BM). Results showed that VOC was gradually increased under continuous illumination of solar light (100 mW/cm2) for ternary solar cells, whereas no VOC increase was measured for binary solar cells without PBDTTPD. As a consequence, the power conversion efficiency (PCE) of ternary solar cells (except the highest PBDTTPD content) was rather higher after solar light illumination for 10 h, even though the binary solar cell exhibited significantly lowered PCE after 10 h illumination. The VOC increase has been attributed to the lateral phase segregation between P3HT and PBDTTPD domains in the ternary BHJ layers under continuous illumination of solar light, as evidenced from the analysis result by Raman spectroscopy, atomic force microscopy, transmission electron microscopy, and synchrotron radiation grazing-incidence angle X-ray diffraction measurements. KEYWORDS: Polymer solar cells, Ternary BHJ, Light-induced, Open circuit voltage, PBDTTPD

■

INTRODUCTION

Although such new conjugated polymers have been recently introduced, regioregular poly(3-hexylthiophene) (P3HT) is still considered one of the most attractive electron-donating polymers because it can deliver 4−5% PCE by making a BHJ layer with soluble fullerene derivatives such as [6,6]-phenylC61-butyric acid methyl ester (PC61BM).14−17 In particular, the P3HT polymer exhibited ∼7% PCE in combination with an indene-C60 bisadduct (ICBA) featuring a lower energy level of a

Polymer solar cells with bulk heterojunction (BHJ) layers of electron-donating conjugated polymers and electron-accepting fullerene derivatives, which are basically solution processable, have been extensively studied because of their potential for cost-effective manufacturing of lightweight and flexible solar modules by employing continuous roll-to-roll processes.1−10 The power conversion efficiency (PCE) of polymer:fullerene solar cells has reached 8−10% due to new conjugated polymers with low band gaps and/or deep highest occupied molecular orbital (HOMO) energy levels.11−13 © 2014 American Chemical Society

Received: August 5, 2014 Revised: November 21, 2014 Published: December 17, 2014 55

dx.doi.org/10.1021/sc500505t | ACS Sustainable Chem. Eng. 2015, 3, 55−62

ACS Sustainable Chemistry & Engineering

Research Article

lowest unoccupied molecular orbital (LUMO) leading to an open circuit voltage (VOC) increase.18,19 However, despite the improved performance of polymer:fullerene solar cells, their stability issue has been still unresolved so that it has been a hurdle for commercialization. The poor device stability can be ascribed to three major reasons: (1) corrosion of transparent electrodes caused by strong acidity of hole collecting buffer layer (HCBL), (2) degradation of conjugated polymers under solar light, and (3) morphological instability in the BHJ layers and/or recrystallization of fullerene derivatives.20−24 For such reasons, the morphological instability, which may be closely related to the gradual demixing (phase segregation) between conjugated polymers and fullerene derivatives, is regarded as the most critical bottleneck for achieving the stability of polymer:fullerene solar cells.24−27 The morphology change by the thermally induced PC61BM crystallization within the P3HT:PC61BM blend films has been reported,25,26 while the light-induced morphology change leading to the PC61BM aggregation has been reported for the P3HT:PC61BM blend film even though the glass transition temperatures of P3HT were considerably higher than typical temperatures of solar cells under solar light.27 In this work, as an alternative approach for the stability improvement in the polymer:fullerene solar cells, we attempted to introduce a third polymer to the P3HT:PC61BM blends in order to investigate the role of the third component (polymer) on the stability of solar cells. As the third component, poly[(4,8-bis(2-ethylhexyloxy)-benzo[1,2-b:4,5-b′]dithiophene)-2,6-diyl-alt-(N-2-ethylhexylthieno[3,4-c]pyrrole4,6-dione)-2,6-diyl]] (PBDTTPD) was used in consideration of its energy band structure because its addition resulted in a significant PCE enhancement in the case of the P3HT:PC61BM solar cells with a low fullerene content (P3HT:PC61BM = 1:0.5 by weight) in our recent report.28 In the present study, unlike the previous work, the weight ratio of polymers (P3HT + PBDTTPD) to PC61BM was fixed as 1:1 by weight because this composition has been confirmed to deliver one of the best PCEs for the P3HT:PC61BM solar cells. The resulting ternary solar cells were examined by measuring the device performances under continuous (10 h) illumination of solar light (100 mW/cm2). As a result, all ternary solar cells exhibited noticeably increased VOC leading to the improved PCE even after the 10 h illumination.

■

Figure 1. (a) Chemical structure of materials used in this work. (b) Illustration of device structure (left) and flat energy band diagram for the device with the ternary BHJ layer (right). Note that the energy unit (eV) and minus sign (−) were omitted. (d) Optical absorption coefficient (α) and (e) photoluminescence spectra (excitation at 505 nm) for binary and ternary bulk heterojunction films coated on quartz substrates. ITO−glass substrates were treated inside a UV−ozone cleaner for 20 min in order to remove any remnant organic residues on the surface of the substrates as well as to make the ITO surface hydrophilic. Then the PEDOT:PSS layer (thickness = 60 nm) was spin-coated on top of the ITO−glass substrates, followed by thermal annealing at 200 °C for 15 min. To make active layers, the binary and ternary blend layers (thickness = 60 nm) were spin-coated on the PEDOT:PSS layers and soft-baked at 60 °C for 15 min. These film-coated samples were transferred to a vacuum chamber in an argon-filled glovebox. After pumping the chamber pressure down to ∼1 × 10−6 Torr, lithium fluoride (LiF, thickness = ∼1 nm) and aluminum (Al, thickness = 80 nm) were sequentially deposited on top of the active layers through a shadow mask. The active area of the devices fabricated in this work was 0.09 cm2. The device structure is shown in Figure 1b (left). Finally, all devices were subject to thermal annealing at 150 °C for 10 min and stored inside the same glovebox before measurements. For the morphology and nanostructure measurements, all samples were prepared in the same way as for the device fabrication, whereas the film samples were prepared by spin-coating on quartz substrates for optical measurements. Measurements. The thicknesses of the thin films were measured using a surface profiler (Alpha Step 200, Tencor Instruments). The optical absorption and photoluminescence spectra of the active layers were measured using a UV−visible spectrometer (Optizen 2120, MECASYS) and a photoluminescence (PL) spectrometer (FS-2, SCINCO), respectively. The Raman spectra of the thin films were measured using a Raman spectrometer (Almega X, Thermo). The surface morphologies of the film samples were measured using an

EXPERIMENTAL SECTION

Materials and Solutions. The regioregular P3HT polymer was purchased from Rieke Metal, Inc. The regioregularity, weight-average molecular weight (MW), and polydispersity index (PDI) of the P3HT polymer were 91%, 6.2 × 104 Da, and 2.3, respectively. The PBDTTPD polymer (MW = 1.3 × 105 Da, PDI = 2.5) was supplied from Solarmer Materials, Inc. PC61BM was received from Nano-C and used without further purification, while the PEDOT:PSS solution (PH500) was purchased from H. C. Starck. The chemical structures of materials are shown in Figure 1a. Binary and ternary blend solutions were prepared using chlorobenzene as a solvent at a solid concentration of 20 mg/mL by varying the ratio of the two polymers at a fixed ratio of polymers (P3HT + PBDTTPD) to PC61BM (P3HT:PBDTTPD:PC61BM = 10−X:X:10 by weight, where X = 0, 2, 4, 6). These binary and ternary blend solutions were vigorously stirred on a magnetic stirring plate for 24 h before spin-coating. Thin film and Device Fabrication. Indium−tin oxide (ITO)coated glass substrates were patterned to make the ITO electrodes by employing photolithography/etching process. The patterned ITO− glass substrates were cleaned using acetone and isopropyl alcohol using an ultrasonic cleaner and dried with a nitrogen flow. The dried 56

dx.doi.org/10.1021/sc500505t | ACS Sustainable Chem. Eng. 2015, 3, 55−62

ACS Sustainable Chemistry & Engineering

Research Article

atomic force microscope (AFM, Nanoscope IIIa, Digital Instruments), while the nanostructures of the film samples were measured using a synchrotron radiation-grazing incidence X-ray diffraction (GIXD) system (X-ray wavelength = 0.11352 nm, incidence angle = 0.12°, 3C SAXS I beamline, Pohang Accelerator Laboratory) and a highresolution transmission electron microscope (HRTEM, HT 7700, Hitachi). The performance of the solar cells under 1 sun condition (100 mW/cm2) was measured using a specialized solar cell measurement system equipped with a solar simulator (92250A-1000, Newport-Oriel) and an electrometer (Model 2400, Keithley). The current density−voltage (J-V) curves of the devices in the dark were measured using the same electrometer in the solar cell measurement system. The short time stabilities of the devices in an argon environment were measured under illumination of solar light (100 mW/cm2) for 10 h.

■

RESULTS AND DISCUSSION Considering the quite large offset (∼0.6 eV) of the lowest unoccupied molecular orbital (LUMO) energy for the three components (see the ideal flat energy band diagram in Figure 1b, right),29,30 the third component (PBDTTPD) is expected to play an intermediate role in the charge separation/transfer process between P3HT and PC61BM components for the present devices with a high fullerene content (P3HT:PC61BM = 1:1 by weight) (note that the LUMO energy value of −4.3 eV was used for PC61BM because most of PC61BM molecules are considered to be aggregated to each other in the binary and ternary films31−34). This assumption is evident from the result that the PL intensity was gradually reduced as the PBDTTPD content increased (Figure 1d). Here, we note that the overall PL intensity should be maintained or increased by the PL of PBDTTPD, which is brighter than P3HT (Figure S1, Supporting Information), if no charge separation process was made and/or any Förster energy transfer from P3HT to PBDTTPD components has occurred (even though the Förster energy transfer seems to be very small when it comes to the band gap energy of the two polymers (Eg = 1.9 eV for P3HT, 1.8 eV for PBDTTPD) from the optical absorption spectra in Figure S2 of the Supporting Information).35−38 Although the PBDTTPD component was expected to ideally play an intermediate role in the charge separation/transfer process, th e p erformance of the t ernary blend (P3HT:PBDTTPD:PC61BM) solar cell was relatively inferior to that of the binary (P3HT:PC61BM) solar cell (Figure S3 and Table S1, Supporting Information), which can be attributed to the relatively poor charge transport for the ternary solar cells as evidenced from their high series resistances under illumination of solar light (Table S1, Supporting Information) and the dark J-V curves (Figure S4, Supporting Information). As shown in Figure 2, a noticeable increase in the open circuit voltage (VOC) was measured for all ternary solar cells under continuous illumination of solar light (100 mW/cm2), even though the VOC value was rather decreased with time for the binary solar cell. We note that only a single work on the VOC increase by post-treatment after device fabrication has been reported for the physical pressing of solar cells with the LiF layer.39 However, the trend of short circuit current density (JSC) was similar for both binary and ternary solar cells. To find any clue of the JSC reduction under continuous illumination of solar light, the color of the active (BHJ) layers was examined, but it was almost not changed for all film samples as shown in the inset photographs in Figure 2. This result may imply that the photoinduced degradation of polymers (P3HT and PBDTTPD) under continuous illumination of solar light is

Figure 2. Light J-V curves for binary and ternary solar cells according to the illumination time under 1 sun conditions (100 mW/cm2): (a) P3HT:PBDTTPD:PC61BM = 10:0:10, (b) P3HT:PBDTTPD:PC61BM = 8:2:10, (c) P3HT:PBDTTPD:PC61BM = 6:4:10, and (d) P3HT:PBDTTPD:PC61BM = 4:6:10. Insets in each graph show the photographs for the binary and ternary BHJ films before and after 10 h illumination.

not a major reason for the JSC reduction, which is supported by the optical properties before and after 10 h illumination (Figures S5 and S6, Supporting Information). To further investigate the performance change of ternary solar cells under continuous illumination of solar light, the detailed trend of solar cell parameters as a function of illumination time was plotted in Figure 3. The VOC value was gradually increased with the illumination time for all ternary solar cells, whereas it was slightly negative for the binary solar cell. In particular, the VOC increase in the ternary (4:6:10) solar cell was ∼0.1 V (from 0.52 to 0.62 V) by 10 h illumination (Table S2, Supporting Information), which corresponds to about a 19% increase in VOC. Here, it is noteworthy that the final VOC value reached 0.62 V for all ternary solar cells, which may indicate the possible limit of VOC in the present device structure upon illumination of solar light. In addition to the VOC increase, the JSC value was also initially (up to 1 h) increased for all ternary solar cells, whereas the binary solar cell exhibited a gradual reduction in JSC. Moreover, the fill factor (FF) was slightly increased with time for the ternary (8:2:10) solar cell and was well maintained for the ternary (6:4:10) solar cell, whereas it was gradually decreased with time for the binary solar cell. As a consequence, the PCE of ternary solar cells (8:2:10 and 6:4:10) was rather higher after 10 h illumination of solar light than that of the fresh devices (note that the maximum PCE was obtained at around 3−5 h for these two ternary solar cells), whereas the PCE of the binary solar cell was gradually decreased with illumination time. To understand the impressive performance increase for the ternary solar cells, the surface morphology of blend films was briefly examined with atomic force microscopy. As shown in Figure 4, the surface morphology of the binary blend film was 57

dx.doi.org/10.1021/sc500505t | ACS Sustainable Chem. Eng. 2015, 3, 55−62

ACS Sustainable Chemistry & Engineering

Research Article

ternary blend film (35% increase). Here, we note that the surface of the ternary blend film was already much coarser than that of the binary blend film before solar light illumination. This reflects that the addition of 10 wt % PBDTTPD did greatly affect the surface morphology of blend films and is considered to be responsible for the gradual VOC increase upon solar light illumination. However, the surface morphology change itself cannot account for the VOC increase in the present ternary solar cells, so we employed a Raman back scattering technique because of its capability for qualitative measurement on the chemical composition in the surface region of thin films.40 As shown in Figure 5, the characteristic Raman peaks for P3HT (1375 cm−1 for the C−C skeletal stretch mode and 1442 cm−1 for the CC symmetric stretch mode in Figure S7 of the Supporting Information)26,40−43 were almost marginally changed after 10 h illumination for the binary blend film, which is in good agreement with the AFM result in Figure 4. In contrast, the ternary (8:2:10) blend film showed noticeable changes in the Raman intensity after 10 h illumination. The representative P3HT peaks at 1375 and 1442 cm−1 were obviously decreased after 10 h illumination, whereas the PBDTTPD peaks at 1477 and 1517 cm−1 were pronouncedly increased. This result does directly indicate that the PBDTTPD composition became relatively enriched in the surface region of the ternary blend film, which further implies that the enriched PBDTTPD component might be the pop-up nanodomains grown after 10 h illumination in the AFM image (Figure 4b, right). This assumption is supported from the TEM image (Figure 6) because the darker domains in Figure 6b are obviously different from the nanoscale morphology of the binary (P3HT:PC61BM) blend film in Figure 6a. Here,

Figure 3. Change of VOC, JSC, FF, and PCE as a function of illumination time for binary and ternary solar cells: (a) linear scale and (b) semilogarithmic scale. The composition for the BHJ layers is given in the top in panel (a) and in the bottom in panel (b): (black circles) P3HT:PBDTTPD:PC61BM = 10:0:10, (red squares) P3HT:PBDTTPD:PC61BM = 8:2:10, (blue triangles) P3HT:PBDTTPD:PC 61 BM = 6:4:10, and (green diamonds) P3HT:PBDTTPD:PC61BM = 4:6:10.

marginally changed, but a more pronounced morphology change was measured for the ternary blend film (8:2:10) after 10 h illumination. This is evident from the different root-meansquare roughness (Rg) changes, such as 0.85 to 0.98 nm (15% increase) for the binary blend film and 2.69 to 3.62 nm for the

Figure 4. 3D height-mode AFM images (2D phase mode: upper right) for (a) binary (P3HT:PBDTTPD:PC61BM = 10:0:10) and (b) ternary (P3HT:PBDTTPD:PC61BM = 8:2:10) BHJ layers before and after 10 h solar light illumination. 58

dx.doi.org/10.1021/sc500505t | ACS Sustainable Chem. Eng. 2015, 3, 55−62

ACS Sustainable Chemistry & Engineering

Research Article

Figure 5. Raman spectra for (top) binary (P3HT:PBDTTPD:PC61BM = 10:0:10) and (bottom) ternary (P3HT:PBDTTPD:PC61BM = 8:2:10) BHJ layers before and after 10 h solar light illumination. The excitation wavelength was 780 nm for all samples.

particular attention is paid to the tiny black spots (